[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Nishanth Menon]

From: Wenbiao Wang <wwang@ti.com>

Voltage Processor state machine transition to disable need to
occur from IDLE state. When we transition OPP in a functioning
system, the call sequence for an OPP transition is as follows:
omap_sr_disable
      -> sr class 3 disable
           -> vp disable
           -> sr disable
forceupdate to voltage/frequency scale depending on which OPP
we are transitioning to.

If we hit a critical timing window where SR had commanded VP
for a voltage transition and VP is in the middle of operating
on that command, it needs to go through a few states before
going to update state(where it actually sends the command to
VC). Initial view of h/w owners is that the state disable of VP
is expected to be sampled for the next transition.

Instead, to be on a safer side, we ensure that the valid states
of the VP state machine is diligently followed by software. This
can be done by waiting for VP to be in idle  prior to disabling
VP. Existing prints have been updated to ensure context is
available on error messages.

As part of this change, increase timeout for VP idle check to
improbable 500uSec to be certain that system is indeed unable
to continue before crashing out with error(worst case expectancy
remains the same 3-100uSec depending on when we caught VP).

Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-omap at vger.kernel.org
Cc: linux-arm-kernel at lists.infradead.org

[nm at ti.com: port from android]
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Wenbiao Wang <wwang@ti.com>
---
V2: fix commit message - picked up redundant info from get_maintainer.pl
Apologies on the spam

V1: http://marc.info/?l=linux-omap&m=133736493721977&w=2

 arch/arm/mach-omap2/vp.c |   15 +++++++++++++--
 arch/arm/mach-omap2/vp.h |    2 +-
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
index f95c1ba..925d869 100644
--- a/arch/arm/mach-omap2/vp.c
+++ b/arch/arm/mach-omap2/vp.c
@@ -262,6 +262,17 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 		return;
 	}
 
+	/*
+	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
+	 * Depending on if we catch VP in the middle of an SR operation.
+	 */
+	omap_test_timeout((voltdm->read(vp->vstatus)),
+			  VP_IDLE_TIMEOUT, timeout);
+
+	if (timeout >= VP_IDLE_TIMEOUT)
+		pr_warning("%s: vdd_%s idle timedout before disable\n",
+			   __func__, voltdm->name);
+
 	/* Disable VP */
 	vpconfig = voltdm->read(vp->vpconfig);
 	vpconfig &= ~vp->common->vpconfig_vpenable;
@@ -274,8 +285,8 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 			  VP_IDLE_TIMEOUT, timeout);
 
 	if (timeout >= VP_IDLE_TIMEOUT)
-		pr_warning("%s: vdd_%s idle timedout\n",
-			__func__, voltdm->name);
+		pr_warning("%s: vdd_%s idle timedout after disable\n",
+			   __func__, voltdm->name);
 
 	vp->enabled = false;
 
diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
index 7c155d2..0abf895 100644
--- a/arch/arm/mach-omap2/vp.h
+++ b/arch/arm/mach-omap2/vp.h
@@ -31,7 +31,7 @@ struct voltagedomain;
 #define OMAP4_VP_VDD_MPU_ID 2
 
 /* XXX document */
-#define VP_IDLE_TIMEOUT		200
+#define VP_IDLE_TIMEOUT		500
 #define VP_TRANXDONE_TIMEOUT	300
 
 /**
-- 
1.7.9.5

[PATCH 0/3] ARM: OMAP3+: VP: collate fixes

[Nishanth Menon]

Reviewing the android code base brought out a few fixes that upstream
does not have, so collating them and posting them out in this chain.

Nishanth Menon (2):
  ARM: OMAP3+: PM: VP: check to ensure VP is idle before forceupdate
  ARM: OMAP3+: PM: VP: check only the VPINIDLE status bit

Wenbiao Wang (1):
  ARM: OMAP3+: PM: VP: ensure VP is idle before disable

 arch/arm/mach-omap2/vp.c          |   34 ++++++++++++++++++++++++++++++----
 arch/arm/mach-omap2/vp.h          |    3 ++-
 arch/arm/mach-omap2/vp3xxx_data.c |    1 +
 arch/arm/mach-omap2/vp44xx_data.c |    1 +
 4 files changed, 34 insertions(+), 5 deletions(-)

Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-omap at vger.kernel.org
Cc: linux-arm-kernel at lists.infradead.org
Regards,
Nishanth Menon
-- 
1.7.9.5

[PATCH 1/3] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Nishanth Menon]

From: Wenbiao Wang <wwang@ti.com>

Voltage Processor state machine transition to disable need to
occur from IDLE state. When we transition OPP in a functioning
system, the call sequence for an OPP transition is as follows:
omap_sr_disable
      -> sr class 3 disable
           -> vp disable
           -> sr disable
forceupdate to voltage/frequency scale depending on which OPP
we are transitioning to.

If we hit a critical timing window where SR had commanded VP
for a voltage transition and VP is in the middle of operating
on that command, it needs to go through a few states before
going to update state(where it actually sends the command to
VC). Initial view of h/w owners is that the state disable of VP
is expected to be sampled for the next transition.

Instead, to be on a safer side, we ensure that the valid states
of the VP state machine is diligently followed by software. This
can be done by waiting for VP to be in idle  prior to disabling
VP. Existing prints have been updated to ensure context is
available on error messages.

As part of this change, increase timeout for VP idle check to
improbable 500uSec to be certain that system is indeed unable
to continue before crashing out with error(worst case expectancy
remains the same 3-100uSec depending on when we caught VP).

Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-omap at vger.kernel.org
Cc: linux-arm-kernel at lists.infradead.org

[nm at ti.com: port from android]
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Wenbiao Wang <wwang@ti.com>
---
 arch/arm/mach-omap2/vp.c |   15 +++++++++++++--
 arch/arm/mach-omap2/vp.h |    2 +-
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
index f95c1ba..925d869 100644
--- a/arch/arm/mach-omap2/vp.c
+++ b/arch/arm/mach-omap2/vp.c
@@ -262,6 +262,17 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 		return;
 	}
 
+	/*
+	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
+	 * Depending on if we catch VP in the middle of an SR operation.
+	 */
+	omap_test_timeout((voltdm->read(vp->vstatus)),
+			  VP_IDLE_TIMEOUT, timeout);
+
+	if (timeout >= VP_IDLE_TIMEOUT)
+		pr_warning("%s: vdd_%s idle timedout before disable\n",
+			   __func__, voltdm->name);
+
 	/* Disable VP */
 	vpconfig = voltdm->read(vp->vpconfig);
 	vpconfig &= ~vp->common->vpconfig_vpenable;
@@ -274,8 +285,8 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 			  VP_IDLE_TIMEOUT, timeout);
 
 	if (timeout >= VP_IDLE_TIMEOUT)
-		pr_warning("%s: vdd_%s idle timedout\n",
-			__func__, voltdm->name);
+		pr_warning("%s: vdd_%s idle timedout after disable\n",
+			   __func__, voltdm->name);
 
 	vp->enabled = false;
 
diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
index 7c155d2..0abf895 100644
--- a/arch/arm/mach-omap2/vp.h
+++ b/arch/arm/mach-omap2/vp.h
@@ -31,7 +31,7 @@ struct voltagedomain;
 #define OMAP4_VP_VDD_MPU_ID 2
 
 /* XXX document */
-#define VP_IDLE_TIMEOUT		200
+#define VP_IDLE_TIMEOUT		500
 #define VP_TRANXDONE_TIMEOUT	300
 
 /**
-- 
1.7.9.5

[PATCH 2/3] ARM: OMAP3+: PM: VP: check to ensure VP is idle before forceupdate

[Nishanth Menon]

Ideally in the flow of DVFS programming, VP should be in idle state
(since we disabled it) before entering forceupdate. Ensure that
this is the case. Not doing this could cause VP statemachine
to enter invalid states.

Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-omap at vger.kernel.org
Cc: linux-arm-kernel at lists.infradead.org

Signed-off-by: Vinay Amancha <vinaykumar@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
---
 arch/arm/mach-omap2/vp.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
index 925d869..985091b 100644
--- a/arch/arm/mach-omap2/vp.c
+++ b/arch/arm/mach-omap2/vp.c
@@ -123,6 +123,18 @@ int omap_vp_forceupdate_scale(struct voltagedomain *voltdm,
 	u8 target_vsel, current_vsel;
 	int ret, timeout = 0;
 
+/*
+	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
+	 * This is an additional allowance to ensure we are in proper state
+	 * to enter into forceupdate state transition.
+	 */
+	omap_test_timeout((voltdm->read(vp->vstatus)),
+			  VP_IDLE_TIMEOUT, timeout);
+
+	if (timeout >= VP_IDLE_TIMEOUT)
+		pr_warning("%s: vdd_%s idle timedout forceupdate(v=%ld)\n",
+			   __func__, voltdm->name, target_volt);
+
 	ret = omap_vc_pre_scale(voltdm, target_volt, &target_vsel, &current_vsel);
 	if (ret)
 		return ret;
-- 
1.7.9.5

[PATCH 3/3] ARM: OMAP3+: PM: VP: check only the VPINIDLE status bit

[Nishanth Menon]

Currently we check against the entire 32 bits of the status register
Where, bits 1-31 are marked as reserved and mentioned in TRM as read
returns undefined values. Hence, check against purely the vpinidle
status bit.

Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-omap at vger.kernel.org
Cc: linux-arm-kernel at lists.infradead.org

Reported-by: Vinay Amancha <vinaykumar@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
---
 arch/arm/mach-omap2/vp.c          |   15 +++++++++------
 arch/arm/mach-omap2/vp.h          |    1 +
 arch/arm/mach-omap2/vp3xxx_data.c |    1 +
 arch/arm/mach-omap2/vp44xx_data.c |    1 +
 4 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
index 985091b..cdcac16 100644
--- a/arch/arm/mach-omap2/vp.c
+++ b/arch/arm/mach-omap2/vp.c
@@ -128,8 +128,9 @@ int omap_vp_forceupdate_scale(struct voltagedomain *voltdm,
 	 * This is an additional allowance to ensure we are in proper state
 	 * to enter into forceupdate state transition.
 	 */
-	omap_test_timeout((voltdm->read(vp->vstatus)),
-			  VP_IDLE_TIMEOUT, timeout);
+	omap_test_timeout((voltdm->read(vp->vstatus) &
+			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
+			  timeout);
 
 	if (timeout >= VP_IDLE_TIMEOUT)
 		pr_warning("%s: vdd_%s idle timedout forceupdate(v=%ld)\n",
@@ -278,8 +279,9 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
 	 * Depending on if we catch VP in the middle of an SR operation.
 	 */
-	omap_test_timeout((voltdm->read(vp->vstatus)),
-			  VP_IDLE_TIMEOUT, timeout);
+	omap_test_timeout((voltdm->read(vp->vstatus) &
+			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
+			  timeout);
 
 	if (timeout >= VP_IDLE_TIMEOUT)
 		pr_warning("%s: vdd_%s idle timedout before disable\n",
@@ -293,8 +295,9 @@ void omap_vp_disable(struct voltagedomain *voltdm)
 	/*
 	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
 	 */
-	omap_test_timeout((voltdm->read(vp->vstatus)),
-			  VP_IDLE_TIMEOUT, timeout);
+	omap_test_timeout((voltdm->read(vp->vstatus) &
+			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
+			  timeout);
 
 	if (timeout >= VP_IDLE_TIMEOUT)
 		pr_warning("%s: vdd_%s idle timedout after disable\n",
diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
index 0abf895..876b5ab 100644
--- a/arch/arm/mach-omap2/vp.h
+++ b/arch/arm/mach-omap2/vp.h
@@ -73,6 +73,7 @@ struct omap_vp_common {
 	u8 vpconfig_initvdd;
 	u8 vpconfig_forceupdate;
 	u8 vpconfig_vpenable;
+	u8 vstatus_vpidle;
 	u8 vstepmin_stepmin_shift;
 	u8 vstepmin_smpswaittimemin_shift;
 	u8 vstepmax_stepmax_shift;
diff --git a/arch/arm/mach-omap2/vp3xxx_data.c b/arch/arm/mach-omap2/vp3xxx_data.c
index bd89f80..a65909b 100644
--- a/arch/arm/mach-omap2/vp3xxx_data.c
+++ b/arch/arm/mach-omap2/vp3xxx_data.c
@@ -44,6 +44,7 @@ static const struct omap_vp_common omap3_vp_common = {
 	.vpconfig_initvdd = OMAP3430_INITVDD_MASK,
 	.vpconfig_forceupdate = OMAP3430_FORCEUPDATE_MASK,
 	.vpconfig_vpenable = OMAP3430_VPENABLE_MASK,
+	.vstatus_vpidle = OMAP3430_VPINIDLE_MASK,
 	.vstepmin_smpswaittimemin_shift = OMAP3430_SMPSWAITTIMEMIN_SHIFT,
 	.vstepmax_smpswaittimemax_shift = OMAP3430_SMPSWAITTIMEMAX_SHIFT,
 	.vstepmin_stepmin_shift = OMAP3430_VSTEPMIN_SHIFT,
diff --git a/arch/arm/mach-omap2/vp44xx_data.c b/arch/arm/mach-omap2/vp44xx_data.c
index 8c031d1..9d14881 100644
--- a/arch/arm/mach-omap2/vp44xx_data.c
+++ b/arch/arm/mach-omap2/vp44xx_data.c
@@ -44,6 +44,7 @@ static const struct omap_vp_common omap4_vp_common = {
 	.vpconfig_initvdd = OMAP4430_INITVDD_MASK,
 	.vpconfig_forceupdate = OMAP4430_FORCEUPDATE_MASK,
 	.vpconfig_vpenable = OMAP4430_VPENABLE_MASK,
+	.vstatus_vpidle = OMAP4430_VPINIDLE_MASK,
 	.vstepmin_smpswaittimemin_shift = OMAP4430_SMPSWAITTIMEMIN_SHIFT,
 	.vstepmax_smpswaittimemax_shift = OMAP4430_SMPSWAITTIMEMAX_SHIFT,
 	.vstepmin_stepmin_shift = OMAP4430_VSTEPMIN_SHIFT,
-- 
1.7.9.5

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Eduardo Valentin]

Hello,

Minor suggestion.

On Fri, May 18, 2012 at 01:18:41PM -0500, ext Nishanth Menon wrote:
> From: Wenbiao Wang <wwang@ti.com>
> 
> Voltage Processor state machine transition to disable need to
> occur from IDLE state. When we transition OPP in a functioning
> system, the call sequence for an OPP transition is as follows:
> omap_sr_disable
>       -> sr class 3 disable
>            -> vp disable
>            -> sr disable
> forceupdate to voltage/frequency scale depending on which OPP
> we are transitioning to.
> 
> If we hit a critical timing window where SR had commanded VP
> for a voltage transition and VP is in the middle of operating
> on that command, it needs to go through a few states before
> going to update state(where it actually sends the command to
> VC). Initial view of h/w owners is that the state disable of VP
> is expected to be sampled for the next transition.
> 
> Instead, to be on a safer side, we ensure that the valid states
> of the VP state machine is diligently followed by software. This
> can be done by waiting for VP to be in idle  prior to disabling
> VP. Existing prints have been updated to ensure context is
> available on error messages.
> 
> As part of this change, increase timeout for VP idle check to
> improbable 500uSec to be certain that system is indeed unable
> to continue before crashing out with error(worst case expectancy
> remains the same 3-100uSec depending on when we caught VP).
> 


<cut>

>  
>  /* XXX document */
> -#define VP_IDLE_TIMEOUT		200
> +#define VP_IDLE_TIMEOUT		500

I guess it is time to properly document this increasing busy loop delay..
As it is getting closer to ms scale..

>  #define VP_TRANXDONE_TIMEOUT	300
>  
>  /**
> -- 
> 1.7.9.5
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Menon, Nishanth]

 Sat, May 19, 2012 at 4:52 AM, Eduardo Valentin <eduardo.valentin@ti.com>
wrote:
>
>
> I guess it is time to properly document this increasing busy loop delay..
> As it is getting closer to ms scale..
Does the following sound good?
/* Maximum time for Voltage Processor to enter or exit idle */

Regards,
Nishanth Menon

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Eduardo Valentin]

Hello,

On Mon, May 21, 2012 at 08:36:35AM -0500, ext Nishanth Menon wrote:
>  Sat, May 19, 2012 at 4:52 AM, Eduardo Valentin <eduardo.valentin@ti.com>
> wrote:
> >
> >
> > I guess it is time to properly document this increasing busy loop delay..
> > As it is getting closer to ms scale..
> Does the following sound good?
> /* Maximum time for Voltage Processor to enter or exit idle */

Sounds way better :-). If you have an estimation of how long it takes
in the average case, it might help. But I am OK already with the above,
in case you don't have the estimation.

> 
> Regards,
> Nishanth Menon

Cheers,

--
Eduardo Valentin

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Menon, Nishanth]

On Mon, May 21, 2012 at 9:51 AM, Eduardo Valentin
<eduardo.valentin@ti.com> wrote:
>
>
> On Mon, May 21, 2012 at 08:36:35AM -0500, ext Nishanth Menon wrote:
>> ?Sat, May 19, 2012 at 4:52 AM, Eduardo Valentin <eduardo.valentin@ti.com>
>> wrote:
>> >
>> >
>> > I guess it is time to properly document this increasing busy loop delay..
>> > As it is getting closer to ms scale..
>> Does the following sound good?
>> /* Maximum time for Voltage Processor to enter or exit idle */
>
> Sounds way better :-). If you have an estimation of how long it takes
> in the average case, it might help. But I am OK already with the above,
> in case you don't have the estimation.
/*
 * Maximum time for Voltage Processor to enter or exit idle
 * worst case is around 100uSec depending on when we intercepted VP
 * we use 5 times worst case value to be sure that the system flags
 * invalid condition
 */
better?

Regards,
Nishanth Menon

[PATCH V2] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Valentin, Eduardo]

Hello,

On Mon, May 21, 2012 at 6:14 PM, Menon, Nishanth <nm@ti.com> wrote:
> On Mon, May 21, 2012 at 9:51 AM, Eduardo Valentin
> <eduardo.valentin@ti.com> wrote:
>>
>>
>> On Mon, May 21, 2012 at 08:36:35AM -0500, ext Nishanth Menon wrote:
>>> ?Sat, May 19, 2012 at 4:52 AM, Eduardo Valentin <eduardo.valentin@ti.com>
>>> wrote:
>>> >
>>> >
>>> > I guess it is time to properly document this increasing busy loop delay..
>>> > As it is getting closer to ms scale..
>>> Does the following sound good?
>>> /* Maximum time for Voltage Processor to enter or exit idle */
>>
>> Sounds way better :-). If you have an estimation of how long it takes
>> in the average case, it might help. But I am OK already with the above,
>> in case you don't have the estimation.
> /*
> ?* Maximum time for Voltage Processor to enter or exit idle
> ?* worst case is around 100uSec depending on when we intercepted VP
> ?* we use 5 times worst case value to be sure that the system flags
> ?* invalid condition
> ?*/
> better?

Yes! Thanks.

>
> Regards,
> Nishanth Menon



-- 

Eduardo Valentin

[PATCH 1/3] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Kevin Hilman]

Nishanth Menon <nm@ti.com> writes:

> From: Wenbiao Wang <wwang@ti.com>
>
> Voltage Processor state machine transition to disable need to
> occur from IDLE state. When we transition OPP in a functioning
> system, the call sequence for an OPP transition is as follows:
> omap_sr_disable
>       -> sr class 3 disable
>            -> vp disable
>            -> sr disable
> forceupdate to voltage/frequency scale depending on which OPP
> we are transitioning to.
>
> If we hit a critical timing window where SR had commanded VP
> for a voltage transition and VP is in the middle of operating
> on that command, it needs to go through a few states before
> going to update state(where it actually sends the command to
> VC). Initial view of h/w owners is that the state disable of VP
> is expected to be sampled for the next transition.
>
> Instead, to be on a safer side, we ensure that the valid states
> of the VP state machine is diligently followed by software. This
> can be done by waiting for VP to be in idle  prior to disabling
> VP. Existing prints have been updated to ensure context is
> available on error messages.
>
> As part of this change, increase timeout for VP idle check to
> improbable 500uSec to be certain that system is indeed unable
> to continue before crashing out with error(worst case expectancy
> remains the same 3-100uSec depending on when we caught VP).
>
> Cc: Tony Lindgren <tony@atomide.com>
> Cc: Kevin Hilman <khilman@ti.com>
> Cc: linux-omap at vger.kernel.org
> Cc: linux-arm-kernel at lists.infradead.org
>
> [nm at ti.com: port from android]
> Signed-off-by: Nishanth Menon <nm@ti.com>
> Signed-off-by: Wenbiao Wang <wwang@ti.com>
> ---
>  arch/arm/mach-omap2/vp.c |   15 +++++++++++++--
>  arch/arm/mach-omap2/vp.h |    2 +-
>  2 files changed, 14 insertions(+), 3 deletions(-)
>
> diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
> index f95c1ba..925d869 100644
> --- a/arch/arm/mach-omap2/vp.c
> +++ b/arch/arm/mach-omap2/vp.c
> @@ -262,6 +262,17 @@ void omap_vp_disable(struct voltagedomain *voltdm)
>  		return;
>  	}
>  
> +	/*
> +	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
> +	 * Depending on if we catch VP in the middle of an SR operation.
> +	 */
> +	omap_test_timeout((voltdm->read(vp->vstatus)),
> +			  VP_IDLE_TIMEOUT, timeout);
> +
> +	if (timeout >= VP_IDLE_TIMEOUT)
> +		pr_warning("%s: vdd_%s idle timedout before disable\n",
> +			   __func__, voltdm->name);
> +

Shouldn't this bail out here instead of continuing to disable VP?

It would be even better to have some error recovery (or at least
clean failure) here rather than just printing a warning and continuing.

>  	/* Disable VP */
>  	vpconfig = voltdm->read(vp->vpconfig);
>  	vpconfig &= ~vp->common->vpconfig_vpenable;
> @@ -274,8 +285,8 @@ void omap_vp_disable(struct voltagedomain *voltdm)
>  			  VP_IDLE_TIMEOUT, timeout);
>  
>  	if (timeout >= VP_IDLE_TIMEOUT)
> -		pr_warning("%s: vdd_%s idle timedout\n",
> -			__func__, voltdm->name);
> +		pr_warning("%s: vdd_%s idle timedout after disable\n",
> +			   __func__, voltdm->name);
>  
>  	vp->enabled = false;
>  
> diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
> index 7c155d2..0abf895 100644
> --- a/arch/arm/mach-omap2/vp.h
> +++ b/arch/arm/mach-omap2/vp.h
> @@ -31,7 +31,7 @@ struct voltagedomain;
>  #define OMAP4_VP_VDD_MPU_ID 2
>  
>  /* XXX document */
> -#define VP_IDLE_TIMEOUT		200
> +#define VP_IDLE_TIMEOUT		500

Were you planning to insert the comment here as suggested by Eduardo?

Kevin

>  #define VP_TRANXDONE_TIMEOUT	300
>  
>  /**

[PATCH 2/3] ARM: OMAP3+: PM: VP: check to ensure VP is idle before forceupdate

[Kevin Hilman]

Nishanth Menon <nm@ti.com> writes:

> Ideally in the flow of DVFS programming, VP should be in idle state
> (since we disabled it) before entering forceupdate. Ensure that
> this is the case. Not doing this could cause VP statemachine
> to enter invalid states.
>
> Cc: Tony Lindgren <tony@atomide.com>
> Cc: Kevin Hilman <khilman@ti.com>
> Cc: linux-omap at vger.kernel.org
> Cc: linux-arm-kernel at lists.infradead.org
>
> Signed-off-by: Vinay Amancha <vinaykumar@ti.com>
> Signed-off-by: Nishanth Menon <nm@ti.com>
> ---
>  arch/arm/mach-omap2/vp.c |   12 ++++++++++++
>  1 file changed, 12 insertions(+)
>
> diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
> index 925d869..985091b 100644
> --- a/arch/arm/mach-omap2/vp.c
> +++ b/arch/arm/mach-omap2/vp.c
> @@ -123,6 +123,18 @@ int omap_vp_forceupdate_scale(struct voltagedomain *voltdm,
>  	u8 target_vsel, current_vsel;
>  	int ret, timeout = 0;
>  
> +/*

missing indent

> +	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
> +	 * This is an additional allowance to ensure we are in proper state
> +	 * to enter into forceupdate state transition.
> +	 */
> +	omap_test_timeout((voltdm->read(vp->vstatus)),
> +			  VP_IDLE_TIMEOUT, timeout);
> +
> +	if (timeout >= VP_IDLE_TIMEOUT)
> +		pr_warning("%s: vdd_%s idle timedout forceupdate(v=%ld)\n",
> +			   __func__, voltdm->name, target_volt);
> +

Again, some clean failure and error recovery should be done here instead
of just a print and bail.

Kevin

>  	ret = omap_vc_pre_scale(voltdm, target_volt, &target_vsel, &current_vsel);
>  	if (ret)
>  		return ret;

[PATCH 3/3] ARM: OMAP3+: PM: VP: check only the VPINIDLE status bit

[Kevin Hilman]

Nishanth Menon <nm@ti.com> writes:

> Currently we check against the entire 32 bits of the status register
> Where, bits 1-31 are marked as reserved and mentioned in TRM as read
> returns undefined values. 

ouch

> Hence, check against purely the vpinidle status bit.
>
> Cc: Tony Lindgren <tony@atomide.com>
> Cc: Kevin Hilman <khilman@ti.com>
> Cc: linux-omap at vger.kernel.org
> Cc: linux-arm-kernel at lists.infradead.org
>
> Reported-by: Vinay Amancha <vinaykumar@ti.com>
> Signed-off-by: Nishanth Menon <nm@ti.com>

Thanks for the fix.

Looking at this combined with the previous patches, since this same
thing is done several places, it looks like we need a helper function
'omap_vp_wait_for_idle()' or similar.  That function will do the check,
wait for timeout, and return an error if failure.  The callers can than
do appropriate error handling upon failure.

Thanks,

Kevin

> ---
>  arch/arm/mach-omap2/vp.c          |   15 +++++++++------
>  arch/arm/mach-omap2/vp.h          |    1 +
>  arch/arm/mach-omap2/vp3xxx_data.c |    1 +
>  arch/arm/mach-omap2/vp44xx_data.c |    1 +
>  4 files changed, 12 insertions(+), 6 deletions(-)
>
> diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
> index 985091b..cdcac16 100644
> --- a/arch/arm/mach-omap2/vp.c
> +++ b/arch/arm/mach-omap2/vp.c
> @@ -128,8 +128,9 @@ int omap_vp_forceupdate_scale(struct voltagedomain *voltdm,
>  	 * This is an additional allowance to ensure we are in proper state
>  	 * to enter into forceupdate state transition.
>  	 */
> -	omap_test_timeout((voltdm->read(vp->vstatus)),
> -			  VP_IDLE_TIMEOUT, timeout);
> +	omap_test_timeout((voltdm->read(vp->vstatus) &
> +			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
> +			  timeout);
>  
>  	if (timeout >= VP_IDLE_TIMEOUT)
>  		pr_warning("%s: vdd_%s idle timedout forceupdate(v=%ld)\n",
> @@ -278,8 +279,9 @@ void omap_vp_disable(struct voltagedomain *voltdm)
>  	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
>  	 * Depending on if we catch VP in the middle of an SR operation.
>  	 */
> -	omap_test_timeout((voltdm->read(vp->vstatus)),
> -			  VP_IDLE_TIMEOUT, timeout);
> +	omap_test_timeout((voltdm->read(vp->vstatus) &
> +			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
> +			  timeout);
>  
>  	if (timeout >= VP_IDLE_TIMEOUT)
>  		pr_warning("%s: vdd_%s idle timedout before disable\n",
> @@ -293,8 +295,9 @@ void omap_vp_disable(struct voltagedomain *voltdm)
>  	/*
>  	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
>  	 */
> -	omap_test_timeout((voltdm->read(vp->vstatus)),
> -			  VP_IDLE_TIMEOUT, timeout);
> +	omap_test_timeout((voltdm->read(vp->vstatus) &
> +			   vp->common->vstatus_vpidle), VP_IDLE_TIMEOUT,
> +			  timeout);
>  
>  	if (timeout >= VP_IDLE_TIMEOUT)
>  		pr_warning("%s: vdd_%s idle timedout after disable\n",
> diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
> index 0abf895..876b5ab 100644
> --- a/arch/arm/mach-omap2/vp.h
> +++ b/arch/arm/mach-omap2/vp.h
> @@ -73,6 +73,7 @@ struct omap_vp_common {
>  	u8 vpconfig_initvdd;
>  	u8 vpconfig_forceupdate;
>  	u8 vpconfig_vpenable;
> +	u8 vstatus_vpidle;
>  	u8 vstepmin_stepmin_shift;
>  	u8 vstepmin_smpswaittimemin_shift;
>  	u8 vstepmax_stepmax_shift;
> diff --git a/arch/arm/mach-omap2/vp3xxx_data.c b/arch/arm/mach-omap2/vp3xxx_data.c
> index bd89f80..a65909b 100644
> --- a/arch/arm/mach-omap2/vp3xxx_data.c
> +++ b/arch/arm/mach-omap2/vp3xxx_data.c
> @@ -44,6 +44,7 @@ static const struct omap_vp_common omap3_vp_common = {
>  	.vpconfig_initvdd = OMAP3430_INITVDD_MASK,
>  	.vpconfig_forceupdate = OMAP3430_FORCEUPDATE_MASK,
>  	.vpconfig_vpenable = OMAP3430_VPENABLE_MASK,
> +	.vstatus_vpidle = OMAP3430_VPINIDLE_MASK,
>  	.vstepmin_smpswaittimemin_shift = OMAP3430_SMPSWAITTIMEMIN_SHIFT,
>  	.vstepmax_smpswaittimemax_shift = OMAP3430_SMPSWAITTIMEMAX_SHIFT,
>  	.vstepmin_stepmin_shift = OMAP3430_VSTEPMIN_SHIFT,
> diff --git a/arch/arm/mach-omap2/vp44xx_data.c b/arch/arm/mach-omap2/vp44xx_data.c
> index 8c031d1..9d14881 100644
> --- a/arch/arm/mach-omap2/vp44xx_data.c
> +++ b/arch/arm/mach-omap2/vp44xx_data.c
> @@ -44,6 +44,7 @@ static const struct omap_vp_common omap4_vp_common = {
>  	.vpconfig_initvdd = OMAP4430_INITVDD_MASK,
>  	.vpconfig_forceupdate = OMAP4430_FORCEUPDATE_MASK,
>  	.vpconfig_vpenable = OMAP4430_VPENABLE_MASK,
> +	.vstatus_vpidle = OMAP4430_VPINIDLE_MASK,
>  	.vstepmin_smpswaittimemin_shift = OMAP4430_SMPSWAITTIMEMIN_SHIFT,
>  	.vstepmax_smpswaittimemax_shift = OMAP4430_SMPSWAITTIMEMAX_SHIFT,
>  	.vstepmin_stepmin_shift = OMAP4430_VSTEPMIN_SHIFT,

[PATCH 1/3] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Kevin Hilman]

On 05/31/2012 04:21 PM, Kevin Hilman wrote:
>> diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
>> >  index 7c155d2..0abf895 100644
>> >  --- a/arch/arm/mach-omap2/vp.h
>> >  +++ b/arch/arm/mach-omap2/vp.h
>> >  @@ -31,7 +31,7 @@ struct voltagedomain;
>> >    #define OMAP4_VP_VDD_MPU_ID 2
>> >
>> >    /* XXX document */
>> >  -#define VP_IDLE_TIMEOUT		200
>> >  +#define VP_IDLE_TIMEOUT		500
> Were you planning to insert the comment here as suggested by Eduardo?

Oops, ignore this comment.  I see you posted a v3 with this added.

Sorry,

Kevin

[PATCH 1/3] ARM: OMAP3+: PM: VP: ensure VP is idle before disable

[Nishanth Menon]

On 16:21-20120531, Kevin Hilman wrote:
> Nishanth Menon <nm@ti.com> writes:
> 
> > From: Wenbiao Wang <wwang@ti.com>
> >
> > Voltage Processor state machine transition to disable need to
> > occur from IDLE state. When we transition OPP in a functioning
> > system, the call sequence for an OPP transition is as follows:
> > omap_sr_disable
> >       -> sr class 3 disable
> >            -> vp disable
> >            -> sr disable
> > forceupdate to voltage/frequency scale depending on which OPP
> > we are transitioning to.
> >
> > If we hit a critical timing window where SR had commanded VP
> > for a voltage transition and VP is in the middle of operating
> > on that command, it needs to go through a few states before
> > going to update state(where it actually sends the command to
> > VC). Initial view of h/w owners is that the state disable of VP
> > is expected to be sampled for the next transition.
> >
> > Instead, to be on a safer side, we ensure that the valid states
> > of the VP state machine is diligently followed by software. This
> > can be done by waiting for VP to be in idle  prior to disabling
> > VP. Existing prints have been updated to ensure context is
> > available on error messages.
> >
> > As part of this change, increase timeout for VP idle check to
> > improbable 500uSec to be certain that system is indeed unable
> > to continue before crashing out with error(worst case expectancy
> > remains the same 3-100uSec depending on when we caught VP).
> >
> > Cc: Tony Lindgren <tony@atomide.com>
> > Cc: Kevin Hilman <khilman@ti.com>
> > Cc: linux-omap at vger.kernel.org
> > Cc: linux-arm-kernel at lists.infradead.org
> >
> > [nm at ti.com: port from android]
> > Signed-off-by: Nishanth Menon <nm@ti.com>
> > Signed-off-by: Wenbiao Wang <wwang@ti.com>
> > ---
> >  arch/arm/mach-omap2/vp.c |   15 +++++++++++++--
> >  arch/arm/mach-omap2/vp.h |    2 +-
> >  2 files changed, 14 insertions(+), 3 deletions(-)
> >
> > diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
> > index f95c1ba..925d869 100644
> > --- a/arch/arm/mach-omap2/vp.c
> > +++ b/arch/arm/mach-omap2/vp.c
> > @@ -262,6 +262,17 @@ void omap_vp_disable(struct voltagedomain *voltdm)
> >  		return;
> >  	}
> >  
> > +	/*
> > +	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
> > +	 * Depending on if we catch VP in the middle of an SR operation.
> > +	 */
> > +	omap_test_timeout((voltdm->read(vp->vstatus)),
> > +			  VP_IDLE_TIMEOUT, timeout);
> > +
> > +	if (timeout >= VP_IDLE_TIMEOUT)
> > +		pr_warning("%s: vdd_%s idle timedout before disable\n",
> > +			   __func__, voltdm->name);
> > +
> 
> Shouldn't this bail out here instead of continuing to disable VP?
yes - but then the entire voltage path is without error handling :(
I think the cleanup should be done independent of this patch - just
doing a return at this point might actually not effective.
> 
> It would be even better to have some error recovery (or at least
> clean failure) here rather than just printing a warning and continuing.
Once VP does not idle, it is an indication that something seriously has
gone wrong in the VP state machine. VP in PRM module is reset only on
cold reset - so even a traditional arch_reset would not help us here.

The intent here is to be completely sure that VP is in idle state before
we trigger a disable. With the cleanup of all error handling flows, we
might be able to flag all the way back up to calling event (a DVFS
event) and may even be able to trigger appropriate recovery mechanism.

> 
> >  	/* Disable VP */
> >  	vpconfig = voltdm->read(vp->vpconfig);
> >  	vpconfig &= ~vp->common->vpconfig_vpenable;
> > @@ -274,8 +285,8 @@ void omap_vp_disable(struct voltagedomain *voltdm)
> >  			  VP_IDLE_TIMEOUT, timeout);
> >  
> >  	if (timeout >= VP_IDLE_TIMEOUT)
> > -		pr_warning("%s: vdd_%s idle timedout\n",
> > -			__func__, voltdm->name);
> > +		pr_warning("%s: vdd_%s idle timedout after disable\n",
> > +			   __func__, voltdm->name);
> >  
> >  	vp->enabled = false;
> >  
> > diff --git a/arch/arm/mach-omap2/vp.h b/arch/arm/mach-omap2/vp.h
> > index 7c155d2..0abf895 100644
> > --- a/arch/arm/mach-omap2/vp.h
> > +++ b/arch/arm/mach-omap2/vp.h
> > @@ -31,7 +31,7 @@ struct voltagedomain;
> >  #define OMAP4_VP_VDD_MPU_ID 2
> >  
> >  /* XXX document */
> > -#define VP_IDLE_TIMEOUT		200
> > +#define VP_IDLE_TIMEOUT		500
> 
> Were you planning to insert the comment here as suggested by Eduardo?
yep. waiting in case we have more comments :)
---
Regards,
Nishanth Menon

[PATCH 2/3] ARM: OMAP3+: PM: VP: check to ensure VP is idle before forceupdate

[Nishanth Menon]

On 16:23-20120531, Kevin Hilman wrote:
> Nishanth Menon <nm@ti.com> writes:
> 
> > Ideally in the flow of DVFS programming, VP should be in idle state
> > (since we disabled it) before entering forceupdate. Ensure that
> > this is the case. Not doing this could cause VP statemachine
> > to enter invalid states.
> >
> > Cc: Tony Lindgren <tony@atomide.com>
> > Cc: Kevin Hilman <khilman@ti.com>
> > Cc: linux-omap at vger.kernel.org
> > Cc: linux-arm-kernel at lists.infradead.org
> >
> > Signed-off-by: Vinay Amancha <vinaykumar@ti.com>
> > Signed-off-by: Nishanth Menon <nm@ti.com>
> > ---
> >  arch/arm/mach-omap2/vp.c |   12 ++++++++++++
> >  1 file changed, 12 insertions(+)
> >
> > diff --git a/arch/arm/mach-omap2/vp.c b/arch/arm/mach-omap2/vp.c
> > index 925d869..985091b 100644
> > --- a/arch/arm/mach-omap2/vp.c
> > +++ b/arch/arm/mach-omap2/vp.c
> > @@ -123,6 +123,18 @@ int omap_vp_forceupdate_scale(struct voltagedomain *voltdm,
> >  	u8 target_vsel, current_vsel;
> >  	int ret, timeout = 0;
> >  
> > +/*
> 
> missing indent
oops. will fix
> 
> > +	 * Wait for VP idle Typical latency is <2us. Maximum latency is ~100us
> > +	 * This is an additional allowance to ensure we are in proper state
> > +	 * to enter into forceupdate state transition.
> > +	 */
> > +	omap_test_timeout((voltdm->read(vp->vstatus)),
> > +			  VP_IDLE_TIMEOUT, timeout);
> > +
> > +	if (timeout >= VP_IDLE_TIMEOUT)
> > +		pr_warning("%s: vdd_%s idle timedout forceupdate(v=%ld)\n",
> > +			   __func__, voltdm->name, target_volt);
> > +
> 
> Again, some clean failure and error recovery should be done here instead
> of just a print and bail.
thx - we should at least refuse to do forceupdate at this point.

-- 
Regards,
Nishanth Menon

[PATCH 3/3] ARM: OMAP3+: PM: VP: check only the VPINIDLE status bit

[Nishanth Menon]

On 16:27-20120531, Kevin Hilman wrote:
> Nishanth Menon <nm@ti.com> writes:
> 
> > Currently we check against the entire 32 bits of the status register
> > Where, bits 1-31 are marked as reserved and mentioned in TRM as read
> > returns undefined values. 
> 
> ouch
> 
> > Hence, check against purely the vpinidle status bit.
> >
> > Cc: Tony Lindgren <tony@atomide.com>
> > Cc: Kevin Hilman <khilman@ti.com>
> > Cc: linux-omap at vger.kernel.org
> > Cc: linux-arm-kernel at lists.infradead.org
> >
> > Reported-by: Vinay Amancha <vinaykumar@ti.com>
> > Signed-off-by: Nishanth Menon <nm@ti.com>
> 
> Thanks for the fix.
> 
> Looking at this combined with the previous patches, since this same
> thing is done several places, it looks like we need a helper function
> 'omap_vp_wait_for_idle()' or similar.  That function will do the check,
> wait for timeout, and return an error if failure.  The callers can than
> do appropriate error handling upon failure.
> 
yep - i think a static inline should do just fine here considering just
a test_timeout is involved.
will post out a new series in a few mins.
-- 
Regards,
Nishanth Menon